Need all your help/opinions... 2nd Rev. CAD models...

[Albigger]

Here are the 3D CAD models I have been working on this week, as a 2nd revision to my aluminum block (can be seen in my other thread - "just to get some responses").

-----I Want To Hear All Your Comments On This------

this one I want to make out of copper, with a thinner base.

base thickness .094" most of the time, down to .060" thinnest.

the thin base part occurs at the bottom of the 'dimples' in the base.

the height of the lower pillars is 1/4", the second tier is an additional 1/4". Any suggestions on changing those?

the orig. block was 1/8" x 1/8" channels/pillars, the new version is 1/16" x 1/16" (.0625" = ~ 1.6mm) Inlet will be 1/2" NPT with 5/8 ID hose barb (for my setup, anyway) and outlets will be 1/4" NPT with 3/8 ID hose barbs. Space is cramped.

Should flow just as good as my current block or better with drilled out NPT barbs

Surface area in the center 20mm x 20mm x 8mm high 'cube' is roughly 22.3 cm^2

Ratio of water to metal over the center (not including the base thickness) is 3:1 What is ideal ratio? I believe it is more metal though?

[Albigger]

more pics...............













As a side note I should be getting my 1600 XP chip sometime after monday to test the aluminum block on. Hopefully all goes well and I can get around to making the revision ASAP.

[UnloadeD]

Well, I don't know enogh for comments or critiques, but I do have a question. Why the square hole for the inlet? I'm sue there is a purpose for it, I just wanna know what it is.

peace.
unloaded

[Volenti]

Originally posted by UnloadeD
Well, I don't know enogh for comments or critiques, but I do have a question. Why the square hole for the inlet? I'm sue there is a purpose for it, I just wanna know what it is.

peace.
unloaded

to focus the jet of water over the area of the block (and colums) directly above the die, I'd assume.

[utabintarbo]

As I understand it (I could be full of sh*t), dimple-like turbulators would be most effective on water flowing laterally, rather than dropping in from above. It might be a lot of machining pain for little gain. But it does increase surface area!

Next, if you are going with a thin (<2-3mm) base plate, structural integrity becomes an issue. Do the "pins" go all the way to the top (no sections)?

BTW, what CAD s/w are you using? SolidWorks?

Bob

[Albigger]

Originally posted by Volenti
to focus the jet of water over the area of the block (and colums) directly above the die, I'd assume.

correct. I had to have clearance for the upper tier of pillars, and I just mad a square to fit around them. I suppose I could open it up to barb size, but my design focuses on increasing water velocity, which the square should help to do. Plus, as Volenti said, to focus it over the center more so as to not have the water just go in and to the side and out the outlets.

Originally posted by utabintarbo
As I understand it (I could be full of sh*t), dimple-like turbulators would be most effective on water flowing laterally, rather than dropping in from above. It might be a lot of machining pain for little gain. But it does increase surface area!

well, I'm not sure about what is more effective with the dimple-like patterns, but here is how I think of it. If you have water flowing laterally across the dimples, there is the possibility of stagnant water (sitting and getting warmer, effectively doing not much) sitting in the dimples, and having the majority of the moving water pass over everything and not mix well with the water sitting in the dimples. With my rather high velocity/high flow block, the incoming water (hopefully) blasts the water right out of the dimples, so as not to let any water stagnate.

Regardless of whether it works well or not, I would REALLY like to incoporate this design over a flat-bottom, as it presents more than 55% surface area gain!!!!

Originally posted by utabintarbo

Next, if you are going with a thin (<2-3mm) base plate, structural integrity becomes an issue. Do the "pins" go all the way to the top (no sections)?

BTW, what CAD s/w are you using? SolidWorks?

Bob

I haven't read up much on the structural design aspect of blocks, so I wasn't sure how thin started to become a problem. If it is less than 2-3 mm as you say, then I have some issues. But I do know this - I want to go thin base, less than .090" (~2.25 mm)
However, I have thick sidewalls around the water channel area, do you think this will be enough to prevent deformation over the cpu die?

The lower pillars contact the cover of the block (no gap, hopefully) and the same should be true of the higher pillars (though most of the top is gone from them). However I'm not sure how completely I will get metal-to-metal shutoff when using the o-ring, it depends how much it compresses.

Is there any remedy for thin base blocks mounting wise? Even I I did make the hold-down to apply pressure to the middle of the cover, the cover attaches to the base at the sides, so it would still be effectively pushing down on the base block on the sides, causing possible warpage over the die. Suggestions here???

oh and good call - i'm using solidworks 2001.

[utabintarbo]

Originally posted by Albigger


well, I'm not sure about what is more effective with the dimple-like patterns, but here is how I think of it. If you have water flowing laterally across the dimples, there is the possibility of stagnant water (sitting and getting warmer, effectively doing not much) sitting in the dimples, and having the majority of the moving water pass over everything and not mix well with the water sitting in the dimples. With my rather high velocity/high flow block, the incoming water (hopefully) blasts the water right out of the dimples, so as not to let any water stagnate.

Regardless of whether it works well or not, I would REALLY like to incoporate this design over a flat-bottom, as it presents more than 55% surface area gain!!!!

Well, the stagnant water thing does not really apply to rounded dimples as pictured, but the surface area gain sure does!

Originally posted by Albigger

I haven't read up much on the structural design aspect of blocks, so I wasn't sure how thin started to become a problem. If it is less than 2-3 mm as you say, then I have some issues. But I do know this - I want to go thin base, less than .090" (~2.25 mm)
However, I have thick sidewalls around the water channel area, do you think this will be enough to prevent deformation over the cpu die?

The lower pillars contact the cover of the block (no gap, hopefully) and the same should be true of the higher pillars (though most of the top is gone from them). However I'm not sure how completely I will get metal-to-metal shutoff when using the o-ring, it depends how much it compresses.

Is there any remedy for thin base blocks mounting wise? Even I I did make the hold-down to apply pressure to the middle of the cover, the cover attaches to the base at the sides, so it would still be effectively pushing down on the base block on the sides, causing possible warpage over the die. Suggestions here???

oh and good call - i'm using solidworks 2001.

If the shorter posts rest against the top, I don't think you will have enough unsupported area to be an issue (though I have no way to prove that). Forget i mentioned it!

And good work with the SolidWorks thing. I'm learning/using that myself.

Bob

[Albigger]

I may be changing to a 3 layer design - 1/2" copper is much more economical to purchase than .625" or .75" thick is. Can anyone direct me to places to purchase?? I've checked copperstock.com, mcmaster.com, and onlinemetals.com. I've searched for more, but no luck.

Solidworks I find pretty easy to use. Its A LOT like Pro-Engineer, which I had to use for one of my classes, but Solidworks seems a little easier to use and less...finicky i guess. I've also used ANVIL express and trying to get some CATIA experience as well... can never know too much, right?

[jaydee]

Originally posted by Albigger
I may be changing to a 3 layer design - 1/2" copper is much more economical to purchase than .625" or .75" thick is. Can anyone direct me to places to purchase?? I've checked copperstock.com, mcmaster.com, and onlinemetals.com. I've searched for more, but no luck.

No luck with what exactly? I get all my Copper and AL at www.onlinemetals.com. What exactly can't you find?

[Albigger]

For my current design i need .625" thick base and .875" thick cover.

I've found bars .625" x 3" or 4" but they are expensive, as are the 1" x 3"

If I go with 3 layers for my block then I can make everything from .500" thick copper, which seems to be much cheaper.

[jaydee]

Originally posted by Albigger
For my current design i need .625" thick base and .875" thick cover.

I've found bars .625" x 3" or 4" but they are expensive, as are the 1" x 3"

If I go with 3 layers for my block then I can make everything from .500" thick copper, which seems to be much cheaper.

Ok, I see. You can't find the stuff for the original design. The easiest way to make stuff to a desired thickness that is hard to get is to mill it to that size from a larger peice with a mill of course I am not sure if that would be cheaper than special ordering it or not, buit that is what I would do. I assume you have or have access to a mill as this design will need one. Looks to me like the Copper is going to be the cheapest part of this project though. That is some serious milling in that thing. Those pins are going to be hard to do and need a fairly expensive endmill (or 2 or 3 depending on how many you break/dull/destroy trying to make this ).

I still havn't figured out your base yet? I am not understanding why you are milling the sides of the block down and leaving the pins stick above it and then covering it up with the top peice. Why not just make the highest set of pins be the same size as the outside walls so you save alot of milling time.

[Albigger]

Yes, I probably will break quite a few endmills I'll try to be careful though.

And I'm not worried about having to mill down the copper to size, its just that anything over 1/2" thick seems to be really expensive to purchase (like, twice the cost or more). That's why I would consider making my block a 3 - layer design where all 3 pieces could be milled out of copper that is 1/2" thick. As my design stands now I would need one peice at least .625" and one .875" thick. I can find these thicknesses, and larger, but they are not too cheap.

I will be milling the block myself so I don't forsee too much manufacturing costs (other than lots of my time).

About your suggestion - is it to leave all the pins as high as the center ones, and leave the sidewalls that high, or just to leave the sidewalls that high? Either way - I want a very contolled flow through the block with not much water contained in the block. Thus if I left the sidewalls high there would be a much larger volume for water to take up and therefore a much slower water velocity in some parts of the block. Sure I could leave a protrusion sticking down out of the cover to counteract this - but that seems to be about the same amount of milling as I have now.


Anyway - still working on this redesign, I will post more CAD pics later tonight or tomorrow maybe....

[jaydee]

Originally posted by Albigger


About your suggestion - is it to leave all the pins as high as the center ones, and leave the sidewalls that high, or just to leave the sidewalls that high? Either way - I want a very contolled flow through the block with not much water contained in the block. Thus if I left the sidewalls high there would be a much larger volume for water to take up and therefore a much slower water velocity in some parts of the block. Sure I could leave a protrusion sticking down out of the cover to counteract this - but that seems to be about the same amount of milling as I have now.


Anyway - still working on this redesign, I will post more CAD pics later tonight or tomorrow maybe....

Yeah I thought about it and studied the pics more last night and you are right to do it as you have it. It is hard for me to to figure out anothers idea without having blue prints in front of me like I am used to.
As for milling the pins have you actually found an endmill that will mill .59" deep at that small size? I been looking but only see then that will go 1/4" to 1/2" deep before they run out of cutting surface (vertically).

[Albigger]

yeah, its sooooo much easier when you have the 3D model or even just the blueprints.

As for the endmill - I haven't looked yet, but I (hope) I can find one to mill 1/2" deep. The .59 dimension is the overall block height, but .09 is the base thickness, so the talles pins are .5" high. I'm working on it now though, and that height might get dropped a little bit.

[jaydee]

Originally posted by Albigger
yeah, its sooooo much easier when you have the 3D model or even just the blueprints.

As for the endmill - I haven't looked yet, but I (hope) I can find one to mill 1/2" deep. The .59 dimension is the overall block height, but .09 is the base thickness, so the talles pins are .5" high. I'm working on it now though, and that height might get dropped a little bit.

Anything under 1/8" is difficult. I think the 1/32" through 1/8" on this page http://www.rlschmitt.com/Standard%20...nd%20Mills.htm all have .5" of vertical cutting area. I am not sure what the quality is but for this size it will be hard to be picky.

[Albigger]

thanks for the link and everything. I will have to check out what I have or can get locally from some places.

[jaydee]

Originally posted by Albigger
thanks for the link and everything. I will have to check out what I have or can get locally from some places.

You should be able to get them locally allthough they may have to order them. I have a great resource here but they have the same hours that I have at work and can never get there when they are open. When you find the right endmill let us know what it is and how well it works.

[Albigger]

question: i've only milled aluminum and steel (and other misc stuff) before. what type of endmill do you recommend for copper? 4 flute?

[jaydee]

Originally posted by Albigger
question: i've only milled aluminum and steel (and other misc stuff) before. what type of endmill do you recommend for copper? 4 flute?

With copper you need at least a tin coated and preferably better endmill IMO. Copper is so dense and soft it causes great pains in milling and drilling. As for the number fo flutes I am not to sure really. I always liked 2 flute for Aluminum because they don't gum up as bad a 3 and 4 flute. I did run a 4 flute tin coated 3/8" end mill into copper and it did real well, but that was 3/8". I kinda think a 1/8" or smaller might gum up real fast in a 4 flute. Someone else can maybe answer that better.

[Albigger]

well, did some more revising, here's what I think it'll look like:



and



The 3 layer design will allow me to move the barbs slightly more apart (though this is not shown in the pics) and should have more controlled flow rates.

I also changed the pillar heights - bottom tier and top tier each .175" long. This should be more reasonable.

Let me know what you think about the 3 piece design.

[Volenti]

I like it

[jaydee]

I like the 3 peice aswell. It should make it a lot easier to tweak the middle peice as you will not have to re-mill the whole top to make adjustment as if I am seeing it right the middle peice is what will be doing most of the flow control. Looking good. You are going to hate milling Copper though!

[Cathar]

I like the 3-piece design. I'm sure I've seen it somewhere else before though, but I just can't put my finger on where...

[Albigger]

Originally posted by Cathar
I like the 3-piece design. I'm sure I've seen it somewhere else before though, but I just can't put my finger on where...

LOL. yeah, i vaguely remember seeing it somewhere....

it was on this really crappy performing block though....

just kidding of couse!


Yeah, I'm not sure now. Masked G, on OCforums suggested completely doing away with the middle piece. I think I will try it both ways to see whats better.

I dont' want to restrict flow outlet (but i can always make the holes in the middle piece bigger) but I want to direct inlet flow right to the center - without letting it spread out too much.

I guess I'll Just have to make some test versions...

Any other thoughts on the block Cathar?

[Cathar]

Originally posted by Albigger
Any other thoughts on the block Cathar?

Your block as it stands doesn't seem to need the middle plate.

I'm not quite sure what you mean about not having the water spread out, 'cos that's what it's going to want to do. As long as you have the inlet barb directly over the place where you want the water to go it'll push the water down fairly "firmly" against where it's pointed. It's not going to go spreading sideways up near where it just leaves the barb by very much at all.

My biggest concern is ensuring an even spread of water flow out from the middle. The water is going to naturally want to head straight for the outlet sides, and not move up and down to the edges unless you create specialised pressure drop regions that forces the water into those locations.

There's still a lot of work for you to do. I'm also concerned about base flex too with nothing really bracing the tops of the pins, but this is dependent on how thin you make your base. It you stay at or above 2mm thickness you should be fine.

There's a few other things crossing my mind concerning the way the pins will affect fluid flow. I tend to agree with the above statement about having the pins rotated by 45 degrees.